[llvm] r251210 - [Support] Add comment to explain why we can't drop NDEBUG.

[Davide Italiano via llvm-commits]

Author: davide
Date: Sat Oct 24 17:15:32 2015
New Revision: 251210

URL: http://llvm.org/viewvc/llvm-project?rev=251210&view=rev
Log:
[Support] Add comment to explain why we can't drop NDEBUG.

Discussed with: Benjamin Kramer.

Modified:
    llvm/trunk/include/llvm/Support/UnicodeCharRanges.h

Modified: llvm/trunk/include/llvm/Support/UnicodeCharRanges.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/Support/UnicodeCharRanges.h?rev=251210&r1=251209&r2=251210&view=diff
==============================================================================
--- llvm/trunk/include/llvm/Support/UnicodeCharRanges.h (original)
+++ llvm/trunk/include/llvm/Support/UnicodeCharRanges.h Sat Oct 24 17:15:32 2015
@@ -51,6 +51,11 @@ public:
   /// the constructor, so it makes sense to create as few UnicodeCharSet
   /// instances per each array of ranges, as possible.
 #ifdef NDEBUG
+
+  // FIXME: This could use constexpr + static_assert. This way we
+  // may get rid of NDEBUG in this header. Unfortunately there are some
+  // problems to get this working with MSVC 2013. Change this when
+  // the support for MSVC 2013 is dropped.
   LLVM_CONSTEXPR UnicodeCharSet(CharRanges Ranges) : Ranges(Ranges) {}
 #else
   UnicodeCharSet(CharRanges Ranges) : Ranges(Ranges) {